Intellectual Property India Publishes Patent Application for System For Deployment Of Two-Way Relay Network With Multi-Agent Deep Reinforcement Learning (madrl/Marl)-Based Beamforming And Method Thereof.

MUMBAI, India, June 22 -- Intellectual Property India has published a patent application (202631026691 A) filed by Indian Institute Of Technology Kharagpur on March 06, 2026, for System For Deployment Of Two-Way Relay Network With Multi-Agent Deep Reinforcement Learning (madrl/Marl)-Based Beamforming And Method Thereof..

Inventors include Soumyasree Bera; Swastik Chakroborty; and Debarati Sen.

The application for the patent was published on June 12, 2026, under issue no. 24/2026.

Abstract: ABSTRACT TITLE: SYSTEM FOR DEPLOYMENT OF TWO-WAY RELAY NETWORK WITH MULTI-AGENT DEEP REINFORCEMENT LEARNING (MADRL/MARL)-BASED BEAMFORMING AND METHOD THEREOF. The present invention provides a communication system comprising at least two communication nodes (U1 , U2) including two transceivers (100, 300) which are configured to engaged in a two-way network via at least one relay node (200) assisted by at least one intelligent reflecting surface IRS node (400). A network controller (NC) is associated with said communication nodes (U1 , U2) enabling transmission of pilot signals therebetween involving the relay node (200) to estimate channels between the nodes. A MARIT implementing module is included at the relay node (200) for computing precoders for beamforming and phases of the IRS for reflection based on the estimated channels, whereby the computed precoder and IRS phases along with the estimated channels are forwarded to the respective transceivers (100, 300) and IRS node (400) using control channel so as to initiate data transmission between the transceivers (100, 300) using thus designed beamforming and IRS reflection parameters resulting improved spectral efficiency and interference mitigation.

Disclaimer: Curated by HT Syndication.